How can you determine a class from a graph using convolutional neural networks?

Question

Python

There is such data where column 1 is depth (X-axis in the first image), column 3 is the measured value of the substance (Y-axis in the second image), column 5 is known classes (substance):

I thought at first to create many images with a value in each point, but it seems too expensive to me.

(Naturally, an image should be fed to the network input)

Answer 1

A typical multivariate classification problem. It is solved either by any classical method (and with two input parameters and three classes - this is probably preferable) from the Scikit-Learn library, for example.
Well, here the author is playing with Keras:
https://blog.stroganov.pro/classification-neural
network/ You can repeat the feat if you really want to.
But why tie convolutional networks here, which are generally meant to work with visualization, I don’t understand. Because of fashion?